Tokedy Python SDK

Official Python library and command-line tool for Tokedy AI compute.

Tokedy™ serves language models through an OpenAI-compatible API. This package is a thin layer over the official openai library: it points the client at Tokedy, reads your Tokedy API key, and adds the tokedy command. Everything else works exactly like the OpenAI SDK.

Website: https://tokedy.ai

Install

pip install tokedy

Python 3.9 or newer.

Quickstart

from tokedy import Tokedy

client = Tokedy()  # reads TOKEDY_API_KEY
resp = client.chat.completions.create(
    model="<model-id>",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Replace <model-id> with a model id from the models list.

Streaming

Pass stream=True to get the answer as it is written:

from tokedy import Tokedy

client = Tokedy()
stream = client.chat.completions.create(
    model="<model-id>",
    messages=[{"role": "user", "content": "Write a haiku about tea."}],
    stream=True,
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
print()

Async

AsyncTokedy has the same interface, with await:

import asyncio

from tokedy import AsyncTokedy


async def main():
    client = AsyncTokedy()
    resp = await client.chat.completions.create(
        model="<model-id>",
        messages=[{"role": "user", "content": "Hello"}],
    )
    print(resp.choices[0].message.content)


asyncio.run(main())

Streaming works the same way: pass stream=True and iterate with async for.

CLI

Installing the package adds a tokedy command.

# One prompt; the answer streams to stdout
tokedy chat "Say hello" --model <model-id>

# Interactive chat; type exit or press Ctrl+C to quit
tokedy chat --model <model-id>

tokedy --version

--model falls back to the TOKEDY_MODEL environment variable. The CLI prints a one-line banner to stderr, so piping stdout (tokedy chat "..." > answer.txt) captures only the answer.

Configuration

Arguments passed to Tokedy(...) win over environment variables. Every other argument (timeout, max_retries, default_headers, http_client, …) is passed to the OpenAI client unchanged.

Getting an API key

Tokedy is available to approved companies. Request access, then once approved sign in and create a key under API keys. The key is shown once; store it as TOKEDY_API_KEY:

export TOKEDY_API_KEY="sk-ty-v1-..."

Support

• Questions: support@tokedy.ai
• Bugs and feature requests: GitHub issues
• Security reports: security@tokedy.ai (see SECURITY.md)

The library sends no telemetry and never logs or prints your API key.

TOKEDY™ is a trademark of Tokedy, Inc.